Long Range Navigation
|
|
This project is a collaboration with the
Life in the Atacama
project at Carnegie Mellon. The goal is to develop the capability for
an autonomous rover to travel one kilometer and estimate its position
to within 5% of the distance traveled. This will allow scientists to
use orbital or descent imagery to direct the rover to far away science
sites and get within 50 meters of the intended site. This is achieved
by the development of sensing hardware (a sun tracker) and position
estimation software that combines information from the sun tracker,
odometry, and onboard cameras to track the robot's motion.
(more)
|
|
Single Cycle Instrument Placement
|
|
Sojourner, Spirit, and Opportunity require lots of operator
intervention in order to navigate to a target selected by a scientist.
With limited communication windows, it can take the rover three days
to move 10 meters and place an instrument against a sample. The
Single Cycle Instrument Placement project is developing tools that
will enable a rover to autonomously navigate to a rock, assess the
rock, plan the motion of the instrument arm, and place the instrument.
(more)
|
|
MER Microscopic Imager Toolkit
|
|
Sprit and Opportunity are carrying microscopic imagers to capture very
high resolution images of rocks and soil. The instrument has a very
narrow depth of field, so as the camera moves toward a target things
go in and out of focus. The MI Toolkit is a set of tools for merging
the sections of the images that are in focus, and in some cases to
produce 3D models of the rocks and soil samples.
(more)
|
|
Past Projects: |
Combining 2D and 3D Tracking
2D tracking relies on the visual appearance of objects, and often
requires some constancy to the appearance as the camera moves. 3D
tracking makes use of shape or position information, but is harder to
acquire. This project was an effort to use a combination of
techniques to improve the robustness and accuracy of visual tracking
using both 2D appearance and 3D position and shape.
|
|
